国立情報学研究所 - ディジタル・シルクロード・プロジェクト
『東洋文庫所蔵』貴重書デジタルアーカイブ

distance above Jalālpur clearly shows that the river's course has for ages set
against it, as already stated above (p. 27). There, too, the volume of water
carried by the Halkiwānī bed at flood time has been clearly demonstrated.

The second difficulty raised as regards the identification of the Kandar Kas
with the praealta fossa of Curtius is equally groundless. Though the left bank
of the torrent bed at its debouchment half a mile above Jalālpur town is shown
by the one-inch map No. 43. H. 6 at 730 feet and hence, as Professor Breloer
assumes, about 30 feet above the level of the river, ocular inspection makes it
certain that the bed near its mouth, fully 1½ miles lower down, must be sufficiently
filled with water from the river to permit the easy launching of boats and
rafts. This is fully supported by the statement received during my stay at
Jalālpur that during the three preceding years the river floods had repeatedly
reached close to the temples and rest house just outside the town. That the
bottom of the Kandar Kas is sandy near its mouth and also where it passes the
town would not, as suggested, have interfered with the assembly of troops on
the wide clay terraces on either bank near and above the town.

That the bend in the river course which Arrian mentions in connexion with
the ἄκρα is at the spur above Jalālpur very much smaller than the one below
Mangla is certain. But it should be remembered that the bend exists, and that
the boldness and mass of the projecting spur as seen from the flat riverine plain
must necessarily increase the impression of the bend made by the river round
its foot for any one not provided with a map. As regards the fourth observation
bearing on the river channel which, unsuspected before, proved so serious an
obstacle after the crossing, it will suffice to point out that the formation of
such minor beds is quite a regular feature in a river such as the Jhēlum passing
over a wide alluvial plain. What has to be kept in view is the fact that
Arrian (v. xiii. 2) distinctly speaks of a river branch and not of a torrent
bed such as the Jabar Nālā at the foot of the Mangla hills. Just because this
torrent bed is a permanent and well-marked feature it could have easily been
observed well before from the height above the supposed starting-point at the
Pothawāla Kas.

The next point concerns the wooded island which screened the place selected
from the crossing. Regarding it we have Curtius's definite statement (VIII. xiii.
15) that it was 'greater than the rest', though Professor Breloer finds it con-
venient to distinguish it as 'the small island' (p. 202). The exceptional advan-
tages which a large wooded island like Admāna must have offered for screening
such protracted and extensive operations as the assembling and embarkation of a
considerable force and its crossing are so obvious that, compared with the evidence
thus afforded by physical features, the fact of this island nowadays being partially